Which Algorithm Should You Use?

Quick Decision Matrix

Detailed Algorithm Analysis

1. Practical Detector - RECOMMENDED

Best for: Most real-world applications

Performance:

• 98%+ accuracy with baseline methods
• Fast execution on all circuit types
• Reliable results across diverse topologies
• Production-ready implementation

Use Cases:

• Circuit design validation
• DFT (Design for Test) analysis
• General testability assessment
• Educational purposes

Example Results:

Priority Encoder: 113 reconvergences (98.3% match with baseline)
Pipelined Multiplier: 7,137 reconvergences (comprehensive detection)

2. Xu & Edirisuriya Research Algorithm

Best for: Research and complex circuits

Performance:

• Highly selective detection
• Research-grade accuracy
• Excellent for complex circuits (pipelined, register-heavy)
• Note: Very selective on simple circuits

Use Cases:

• Academic research
• Complex VLSI designs
• Pipelined architectures
• Register-heavy circuits

Example Results:

Priority Encoder: 0 reconvergences (too selective for simple circuits)
Pipelined Multiplier: 74 reconvergences (excellent for complex circuits)

3. Baseline BFS Method 📊

Best for: Comparison and legacy compatibility

Performance:

• Traditional approach (well-understood)
• Reliable baseline for comparison
• Good for simple circuits
• Note: May miss complex patterns

Use Cases:

• Algorithm comparison
• Legacy system compatibility
• Educational baseline
• Simple circuit analysis

Circuit-Specific Recommendations

Simple Combinational Circuits

• Primary: Practical Detector
• Alternative: Baseline BFS
• Avoid: Research Algorithm (too selective)

Complex Pipelined Circuits

• Primary: Research Algorithm
• Alternative: Practical Detector
• Avoid: Baseline BFS (may miss patterns)

Register-Heavy Designs

• Primary: Research Algorithm
• Alternative: Practical Detector
• Avoid: Baseline BFS (limited register awareness)

Mixed-Signal or Analog-Digital

• Primary: Practical Detector
• Alternative: Compare all three
• Note: May need custom analysis

Performance Comparison

Getting Started

For New Users

# Start with the practical detector (recommended)
./opentestability practical your_circuit_dag.json

# If you need research-grade analysis
./opentestability research your_circuit_dag.json

# Compare all algorithms to understand differences
./opentestability compare your_circuit_dag.json

For Researchers

# Use research algorithm for academic work
./opentestability research your_circuit_dag.json

# Compare with baseline for validation
./opentestability compare your_circuit_dag.json

For Production Use

# Use practical detector for reliable results
./opentestability practical your_circuit_dag.json

# Validate with comparison if needed
./opentestability compare your_circuit_dag.json

Understanding Results

Reconvergence Counts

• Higher counts: More complex fanout structures detected
• Lower counts: Simpler circuit topology or selective detection
• Zero counts: No reconvergent fanout detected (may indicate simple circuit or very selective algorithm)

Accuracy Assessment

• Practical Detector: Should match baseline within 2-5%
• Research Algorithm: May be significantly different, especially on simple circuits
• Baseline BFS: Traditional approach, good for comparison

Troubleshooting

If Practical Detector Returns 0 Reconvergences

• Check if circuit has fanout points (nodes with multiple outputs)
• Verify DAG construction was successful
• Try baseline BFS for comparison

If Research Algorithm Returns 0 Reconvergences

• This is normal for simple circuits
• Try practical detector for more comprehensive results
• Consider if circuit has complex pipelined structures

If All Algorithms Return 0 Reconvergences

• Circuit may not have reconvergent fanout
• Check circuit topology
• Verify input file format

Advanced Usage

Custom Analysis Pipeline

# Parse and build DAG
./opentestability parse circuit.v circuit --json
./opentestability dag circuit.json

# Run multiple analyses
./opentestability practical circuit_dag.json
./opentestability research circuit_dag.json
./opentestability compare circuit_dag.json

# Generate visualizations
./opentestability visualize circuit_dag.json

Batch Processing

# Process multiple circuits
for circuit in *.v; do
    base=$(basename $circuit .v)
    ./opentestability parse $circuit $base --json
    ./opentestability dag ${base}.json
    ./opentestability practical ${base}_dag.json
done

Need Help?

• Documentation: Check the main README.md
• Issues: Report problems on GitHub Issues
• Discussions: Ask questions on GitHub Discussions

Remember: Start with the Practical Detector for most applications, and use the Research Algorithm for complex circuits or academic work! 🎯